Re: How to shorten output of a hash function?

Phil Carmody <thefatphil_demunged@xxxxxxxxxxx> writes:

Kristian Gjøsteen <kristiag+news@xxxxxxxxxxxx> writes:
Phil Carmody <thefatphil_demunged@xxxxxxxxxxx> wrote:
If it doesn't appear to be a PRF it's not a hash, IMHO.

That opinion is wrong. Sorry.

(Ok, I'll explain: VSH is clearly a hash function. Chaum-van
Heijst-Pfitzmann as well. Neither look very much like random
functions.)

I've got to admit that, on princple, I don't like the idea of
anything which has a <<O(2^(n/2)) work factor to find collisions
for an n-bit hash to be worthy of the title 'n-bit hash'.

hashes were invented for very different purposes than crypto. There are
many hashes which are not cryptographic and which can easily be broken, but
are extermely useful. I would agree that a cryptographic hash should look a
lot like a random function. However, one migh imagine padding a hash (eg
the DES hash) to bring it up to some length. (eg the DES based hash in the
unix password is 13 bytes, but only has a strength of 56 bits, which is 7
bytes. )



Very Sub-exponential Hash, I think you'll find...

What is a sub-exponential hash?


Phil
--
The fact that a believer is happier than a sceptic is no more to the
point than the fact that a drunken man is happier than a sober one.
The happiness of credulity is a cheap and dangerous quality.
-- George Bernard Shaw (1856-1950), Preface to Androcles and the Lion
.

Relevant Pages

  • Re: Potential DoS on Bittorrent
    ... >to the suggestion that anyone who chooses a crypto hash for data integrity ... hash is good choice for a total message integrity check. ... crypto hashes exhibit more graceful degradation in case ... hash functions defined by taking the remainder modulo a composite does ...
    (sci.crypt)
  • Re: bootstrapping a secure channel
    ... if you are trying to verify a cryptographic parameter ... Let's say I have a crypto system and in order to break it you have to ... but requiring still makes it twice as hard ... If I had required the attackers of the SHA-0 hash to get all three (A, ...
    (sci.crypt)
  • Re: Fast 32-bit Hash
    ... > one of those found with 100% certainty). ... CRC's are almost the ideal hash, ... Surely we would not put a 16-bit CRC into a new ... Crypto Glossary http://www.ciphersbyritter.com/GLOSSARY.HTM ...
    (sci.crypt)
  • Re: Sony unveils next generation hash function
    ... Some programmer took a night school course on crypto and felt he could ... Technology) invited entries for the "SHA-3 competition," a selection ... process for next generation cryptographic hash functions. ... candidates, ...
    (sci.crypt)
  • Crypto Update for 2.6.27
    ... Here is the crypto update for 2.6.27: ... The biggest change is the introduction of the new hash interface ... of new hardware support and a lot of hifn fixes. ...
    (Linux-Kernel)
Quantcast